ammonia is composed of azote and hydrogen, than we can that sugar is composed of carbonic oxide and hydrogen, or of carbon and water; because the very instant that these last substances are formed the organic compound ceases to exist as such.

In the present state of our knowledge we must consider ammonia as the oxide of a compound radicle; and the formula expressing the composition of the radicle will be N+ 3 H, and that for ammonia N + O. It is easy to see that by this explanation the phenomena presented by ammonia will cease to be anomalous, and that the analogy of ammonia with other bodies is perfectly restored.

We have, then, two combustible bodies, which, by uniting with hydrogen, produce compound and oxidable radicles; namely, nitric and carbon. A radicle composed of nitric and hydrogen constitutes the base of ammonia; a radicle composed of carbon and hydrogen constitutes the principal base of vegeta ble bodies; a triple radicle composed of carbon, nitric, and hydrogen, constitutes the base of animal bodies. It is by varying on the one hand the proportions of the constituents of the radicles, and on the other that of the degrees of oxidation, that nature with so few elementary principles produces that prodigious diversity of organic products.

What, then, is the amalgam produced by the decomposition of ammonia? Certainly not the ammonia itself combined with hydrogen gas and mercury, as Thenard and Gay-Lussac have concluded from experiments far indeed from being conclusive:* nor is there any other way of explaining this phenomenon, at once singular and perfectly analogous to what happens with the other alkalies, than to consider the amalgam as composed of mercury, and the radicle either of ammonia or of azote, though the last is the least probable. If, on the other side, it is the radicle of ammonia which is combined with mercury, we have the curious phenomenon of a compound metallic body.

Before leaving this subject I trust I shall be permitted to draw the attention of the reader to some circumstances relative to ammonia and its compounds, which are well deserving of consideration:

1. How comes it that nitric, a body so strongly electro

It was taken for granted in these experiments that an ammoniacal amalgam, formed in liquid ammonia, but afterwards fixed by a kind of congelation, ought to contain no water. It is easy to see how little force this argument of Gay-Lussac and Thenard ought to have, when mercury itself in ordinary circumstances cannot be deprived of all the water with which it is penetrated without being boiled. Yet in this amalgam, formed in the midst of water, and ten times less dense than mercury, a dryness so absolute is supposed to exist that it does not coutain water enough to oxidize the infinitely small quantity of metal combined with the mercury.

negative, produces with hydrogen, which is feebly electropositive, a radicle strongly electro-positive? It will be said, because the hydrogen in ammonia is combined with less oxygen than in water, and because Thenard and Gay-Lussac have observed that, when in organic compounds, the hydrogen is to the oxygen as in water, the compounds are neither acid nor alkaline; but when the quantity of oxygen exceeds this ratio the substances are acids. The observation of the French chemists is interesting, but it proves nothing; for in acetic acid, which is one of the strongest vegetable acids, the oxygen is to the hydrogen precisely in the same ratio as in water; and the sulphuret and telluret of hydrogen preserve, in spite of the presence of hydrogen and the absence of oxygen, the electro-negative nature of sulphur and tellurium; so that hydrogen does not appear of itself to determine any thing respecting the electro-chemical nature of a compound. There remains, therefore, something at the bottom of this fact which we do not understand.

2. Water is composed of 11.75 hydrogen and 88.25 oxygen, and nitric acid likewise of 11.71 nitric and 88.29 oxygen. In azote the nitric is combined with 14 times as much oxygen as in ammonia is combined with the hydrogen and nitric toge ther; that is to say, that according to the estimations which appear at present the most correct, the weight of the hydrogen in ammonia is precisely one-half of that of the nitric. Is this merely accidental? or is there some connection of cause which occasions it? Can we suppose that nature, in modifying the electro-chemical state of the radicle of ammonia, produces from it, under different and unknown circumstances, sometimes hydrogen, sometimes the suboxide of nitric? Is it possible that the nitric, that is, the electro-chemical modification which renders it capable of forming an acid, never exists without oxygen; and that on that account azote, though a compound, baffles all our attempts to reduce it: and consequently the combustible radicle can never exist alone under any other forms than. those of ammonium and hydrogen? If these ideas should one day become probable, what views would they not furnish us with respecting the regeneration of the atmosphere, and the production of azote among herbivorous animals, which find so little of it in their food, and yet furnish daily so great a quantity of it in their excretions? But it is easy to make conjectures. Perhaps I have already indulged in them too far. Deceitful probabilities are almost always more injurious to science than the advancement of absurdities, or inaccurate experiments.

ARTICLE IX.

Memoir on the Determination of the Specific Heat of the different Gases. By MM. F. Delaroche, M. D. and J. E. Berard.

Determination of the Specific Heat of the Gases by another Process.

WE might have repeated a greater number of times the experiments which led us to the foregoing results; but we conceived that such a repetition would have been useless. The agreement which exists between those of the same kind, and which we obtained in two experiments following each other, shows that they were not influenced by accidental causes of error; and if they were subject to a constant cause of error, no advantage would result from multiplying their number. It was therefore more essential to endeavour to arrive at the same results by a different process, susceptible of almost equal precision, and to compare these results with the preceding. Count Rumford has put this in our power, by publishing the ingenious method by which he determined the quantity of heat disengaged during the combustion of certain substances, and which he had himself announced as fit for determining the specific heats of the gases.

We therefore undertook some experiments of this kind, employing the same apparatus which has been already described, and which had been constructed long before we were acquainted with the experiments of Count Rumford. It answered all the conditions demanded by this celebrated philosopher. It was, as has been already observed, a metallic vessel full of distilled water, traversed by a flat serpentine, sufficiently long to allow the gas while passing through it to acquire the temperature of the water in the calorimeter. As the gas entered by the lower end, it deposited there the greatest part of its heat; so that this heat spread itself sufficiently equably through the whole water that filled the vessel. We determined the temperature of this water by means of a thermometer, the cylindrical bulb of which was almost as high as the calorimeter itself, and which gave us, of course, the mean heat of the whole liquid. Our gazometer, and the part of the apparatus for heating the gas, were equally fitted for the experiment.

Leaving every thing disposed as in the preceding experiments, we made a new set, proceeding in the same manner, with this difference, that instead of waiting till the temperature of the VOL. II. N° V. 2 A

calorimeter was become stationary, we determined the quantity of gas which passed through the serpentine while that temperature rose a certain given number of degrees, setting out from a given point, and likewise the time necessary to produce this effect. We employed the results of these experiments to calculate the specific heats of the gases experimented upon, setting out from this simple principle, that the specific heats, cæteris paribus, must be inversely as the quantities of gas necessary to produce the same elevation of temperature. In order to make accurate experiments according to this principle, it was necessary to take certain precautions, and to make some corrections on the results, which it will be proper to point out.

Philosophers have long been of opinion that the quantity of heat disengaged during the combustion of different substances. might be determined in this manner; but it was necessary to contend with a cause of error, which proved very injurious to the accuracy of the results. In proportion as the calorimeter became hot, the air and the surrounding bodies deprived it of a portion of its heat. Hence the whole heat of the burning body was not indicated by the rise in the temperature of the calorimeter. It was necessary to include the heat lost during the process, and this was very difficult.

In order to get rid of this inconvenience, Count Rumford conceived the happy thought of beginning the experiment not at the temperature of the surrounding air, but a little below it, and to allow the experiment to continue till the calorimeter was heated as much above the surrounding air, as it had been below it at the commencement. By this method he made the heating of the calorimeter almost independent of the surrounding air. During the first part of the process the air would communicate heat to the calorimeter: but during the second part the calorimeter in its turn would give out nearly as much heat to the air as it had received from it at first.

According to this principle, we sunk the temperature of our calorimeter about 10° Fahrenheit below that of the air of the room; and causing a current of hot gas to pass through it, we did not begin our experiment till the temperature of the calorimeter was within 4° of that of the air of the room. We then began to reckon the number of cubic inches of gas necessary to raise the temperature 8° higher than the initial temperature. By this method we made two experiments at once; one which gave us the quantity of gas necessary to raise the calorimeter 8°; another which gave that which was necessary to raise it the 4 intermediate degrees. It is difficult to say which of these determinations led us to the most exact conclusions; the difference between them was never considerable, as may be seen by the table of our results.

The principle upon which these experiments depended was such that it would have been requisite, in order to be applicable to the object in view, that the gases during their passage through the calorimeter should suffer the same diminution of temperature. As this could not actually take place exactly, we have always by the rule of proportion brought the results obtained to what they would have been if that had taken place. We took for a measure of the cooling of the gases the difference between the temperature of the gases when they entered the calorimeter, and their temperature when they issued out, which we considered as equal to that of the surrounding air. The last estimate could not differ from the truth; for since the gas in passing through the calorimeter assumed the temperature of the water in that vessel, which during one-half of the experiment was below the temperature of the air, and during the other half as much above it, a compensation took place, which enabled us to take the temperature of the air as the mean heat of the gas during the whole process.

The same compensation corrected another error, which might have been occasioned by the condensation of aqueous vapour. The gas saturated with vapour at the temperature of the air would deposite a portion of it in its passage through the serpentine during the first half of the experiment; but as it would take up nearly the same quantity during the second half of the experiment, the general result would not be affected.

We have observed already that we could not entirely prevent the gas tube from communicating some heat to the calorimeter. It was necessary to estimate the influence of this cause of error. We ascertained, by two very correct experiments, that the quantity of heat yielded in this manner to the calorimeter in ten minutes was capable of raising its temperature 0.342° Fahrenheit; and as it was sensibly proportional to the time, nothing was more easy than to allow for it.

We now present the result of our experiments in the following table, quite similar to the preceding one, and therefore requiring no explanation :